Abstract
The collapse of the coral-dinoflagellate relationship under stress, as, for example, induced by increasing sea surface temperatures due to climate change, leads to coral bleaching and coral mortality. While symbiont shuffling or community shifting has been put forth as a rapid adaptive mechanism in corals, reported instances of these phenomena typically focus on environmental extremes rather than natural seasonal increases in sea surface temperatures that may lead to thermal stress, requiring regulation and acclimation of endosymbiotic Symbiodiniaceae. Understanding the dynamic nature of Symbiodiniaceae endosymbiont community responses to seasonal environmental fluctuations is necessary to help predict the limits of acclimation and adaptation potential. We used a combination of flow cytometry, 3D scanning, and ITS2 DNA metabarcoding to quantify Acropora pulchra Symbiodiniaceae community assemblage composition and function in situ in Guam (Micronesia). Samples were collected during the onset of seasonal warming and the time of year during which corals experience the highest sea water temperatures. Flow cytometry allowed us to expediently generate physiological profiles for thousands of individual endosymbiont cells using their autofluorescent signatures. Under variable environmental conditions, Symbiodiniaceae assemblages displayed site and season-specific photophysiological acclimation signatures while community composition and cell densities in host tissues remained homogeneous across sites. Variable photoacclimation patterns during the early season was followed by an island-wide convergence of photophysiological acclimation signatures during the season that sees the highest water temperatures in Guam. Our results show that photoacclimation rather than symbiont community assemblage reorganization allows for acclimation of Acropora pulchra to seasonal extremes in water temperatures.
Introduction
The ecological success of reef-building corals (Scleractinia) can be attributed to the endosymbiotic relationship between corals and their dinoflagellate endosymbionts (Symbiodiniaceae), a relationship that has evolved independently multiple times (Gault et al. 2021). Corals depend on these micro-algal endosymbionts for nutrient acquisition and effective calcification (Roth 2014), but environmental stress can lead to a functional breakdown of this photo-endosymbiotic relationship, leading to the expulsion of Symbiodiniaceae from the host (coral bleaching) (Brown 1997). Climate change has increased the frequency and severity of coral bleaching globally (Hughes et al. 2017). The survival of corals in stressful environments can be linked to differences in the ecological tolerances of Symbiodiniaceae (Parkinson and Coffroth 2015; Thornhill et al. 2014). If the environment shifts, and endosymbionts are not functionally advantageous, successful acclimation is commonly attributed to shifts in endosymbiont community composition, known as symbiont shuffling (Jones et al. 2008; Baker 2003; Buddemeier and Fautin 1993). However, Symbiodiniaceae community composition can be remarkably stable (Rouzé et al. 2019) suggesting alternative modes of acclimation, such as the upregulation of oxidative stress-reducing proteins (Maor-Landaw and Levy 2016).
Despite a large body of research on Symbiodiniaceae function and diversity, little is known about Symbiodiniaceae photophysiological acclimation in situ. Dinoflagellates can reorganize their photopigments and adjust morphology to acclimate to different light conditions (Johnsen et al. 1994; Xiang, 2015). However, prolonged light stress can lead to persistent damage of the Symbiodiniaceae photosystem (Berg et al. 2021). While coral bleaching has been proposed as a mechanism to survive unpredictable change (Baker 2001), Symbiodiniaceae experience no decline in photopigment abundance during coral bleaching (Venn et al. 2006; Jeffrey & Humphrey. 1975). In recent decades, researchers have largely focused on understanding coral endosymbiont dynamics under extreme environmental stress, such as coral bleaching, thereby creating a potential knowledge gap in Symbiodiniaceae ecological tolerances and acclimation dynamics.
To elucidate the scales and mechanisms of Symbiodiniaceae acclimation in corals under natural conditions, we quantified the biodiversity, density, and photophysiology of Symbiodiniaceae in Acropora pulchra at the end of the cooler dry season and in the middle of the hotter wet season. A. pulchra is a dominant ecosystem engineer of Indo-Pacific and Pacific reef flat communities and has been described as having a ‘flexible’ Symbiodiniaceae assemblage (Rouzé et al. 2017). However, we found that Symbiodiniaceae community composition and cell densities in tissues were stable across space and time, while photophysiology varied with site and season suggesting within-cell photopigment regulation as the main mode of seasonal acclimation.
Materials and Methods
Field Sites
To identify the dynamics of Symbiodiniaceae assemblages under natural seasonal fluctuation, four thickets of A. pulchra from five reef flats (20 thickets total) were GPS-tagged and sampled around the island of Guam (Micronesia): Urunao (N 13.63672° E 144.84527°), West Agaña (N 13.47993° E 144.74278°), Luminao (N 13.46584° E 144.64496°), Togcha (N 13.36865° E 144.774967°), and Cocos Lagoon (N 13.24596° E 144.68475°) (Figure 1). A. pulchra was identified using a combination of population genetics (Rios 2020) and corallite morphology. Guam has distinct windward (East) and leeward sides (West). Urunao (North) is a pristine, shallow site with dense thickets of A. pulchra evenly distributed from the algal-dominated reef crest to the beach. West Agaña (Northwest) has a large community of A. pulchra exposed to moderate wave energy. Luminao (West) has a small, scattered population of reasonably large thickets nested within an extensive Porites-dominated reef with a gentle, steady current driven by strong wave energy on a distant algal ridge. Cocos Lagoon (South) has a few small thickets of A. pulchra growing within the lagoon surrounded by sediment and massive Porites or within extensive, shallow water Acropora aspera thickets near the edge of the lagoon. The only known east-coast population of A. pulchra, Togcha (East), is limited to a few remaining thickets in the high wave energy zone near the reef crest.
a) Sampling sites in Guam (scale: 20 km) with each square indicating the location of a reef flat sampled for this study: b) Urunao (N), c) West Agaña (NW), d) Luminao (W), e) Cocos Lagoon (S), and f) Togcha (E). b-f) Within each site, four thickets (1-4) of A. pulchra were georeferenced for repeated sampling. Images provided by Google Earth Pro v7.3.4.8248. (scale: 100 m)
Environmental Data
To characterize seasonal environmental changes, island-wide average sea surface temperature (SST) (NOAA Coral Reef Watch 2019) and precipitation (Menne et al. 2012a; Menne et al. 2012b) were obtained for 2021; bleaching risk was assessed using NOAA’s virtual field station for Guam (Liu et al. 2018). Additionally, historical wave height, period, and cardinal direction were downloaded for 2012-2022 from wave buoys at Ritidian Point (N Guam) and Ipan (SE Guam), provided by the Pacific Islands Ocean Observing System (PacIOOS; www.pacioos.org).
Tissue Sampling
Coral tissue samples were collected at least three centimeters below the axial growth tip, flash-frozen in the field with liquid nitrogen, and stored at -80 °C until further processing. Three technical replicates per colony were collected from 20 tagged colonies during two time periods: 30 April - 18 May (1) and 28 July - 15 August 2021 (2).
Symbiodiniaceae biodiversity
Genomic DNA was extracted using a Qiagen DNeasy PowerSoil Pro Kit (Qiagen, Hilden, Germany) on a Qiacube connect liquid handling system. The ITS2 region was amplified via PCR with SYM_VAR_5.8S2 and SYM_VAR_REV primers (Hume et al. 2018) using 3 μL of DNA (10 ng/μl), 3 μl of 10 μM primer, 2.4 μl of 2.5 mM dNTP, 18.6 μl water, 3 μl buffer (10x), and 0.15 units Taq (TaKaRa Taq™ DNA Polymerase 1U). The PCR profile was 26 cycles of 95°C for 40 s, 59°C for 120 s, 72°C for 60 s, and a final extension at 72°C for 420 s. ITS2 amplicons were multiplexed and sequenced on a NovaSeq 6000 (250bp Paired-end; Illumina, San Diego, CA, USA). Paired-end reads were initially analysed using the SymPortal pipeline with default parameters (Hume et al. 2019) to identify Symbiodiniaceae lineages and assign ITS2 type profiles.
Symbiodiniaceae density
Coral tissue was airbrushed from the skeleton with filtered seawater (FSW) and homogenised using a vortexer and syringe needle-shearing. One ml of tissue homogenate was transferred to 1.5 ml tubes and the remaining volume of tissue homogenate was measured with a graduated cylinder. Modified from the cytometry method described in Krediet et al. (2015), tissue homogenates were vigorously shaken for five seconds using a MiniBeadBeater Plus to separate Symbiodiniaceae cells from host cells and mucus, then centrifuged at 5000 rpm for four minutes at 10 °C. The supernatant was discarded, and the pellet of Symbiodinaceae cells was resuspended in 1 ml of FSW. Shaking (5 s) and centrifugation (3 min) was repeated. The supernatant was discarded and the pellet was resuspended in 1 ml of sodium dodecyl sulphate (SDS) solution (0.08% w/v SDS, 7/8 deionized water (DI), 1/8 FSW). Samples were diluted 1:10 with a 50:50 DI:FSW solution (Total volume: 200 μl) for flow cytometry. Symbiodinaceae concentrations (cells/ml) were estimated using two technical replicates with a Luminex Guava easyCyte 6HT-2L flow cytometer (Luminex Corporation, Austin, TX). Symbiodiniaceae were identified using red fluorescent emission (695±50 nm) from blue light excitation (488 nm laser) and side scatter (Krediet et al. 2015). Absolute cell counts were obtained by multiplying cytometry-generated cell concentrations by each sample’s dilution factor and tissue homogenate volume to determine total cell count for each coral tissue fragment. Cell densities per cm2 (Equation 1) were obtained by the normalization of cell counts to the skeletal surface area. To generate a 3D model for each coral fragment, point clouds with 0.010 mm point spacing were generated using a 3D scanner (D3D-s, Vyshneve, Ukraine). Coral fragments were coated with SKD-S2 Aerosol (Magnaflux, Glenview, IL) to reduce skeletal light refraction. Point clouds of each fragment were imported into MeshLab v2020.04 (Cignoni et al. 2008) to generate a surface mesh by Poisson surface reconstruction. Portions of the fragment that were not covered in tissue prior to airbrushing were removed from the reconstructed surface prior to surface area estimation.
Symbiodiniaceae fluorescent signatures
To characterize the physiological profile of individual dinoflagellate cells, we developed a protocol that quantifies Symbiodiniaceae photosystem components while counting cells using flow cytometry (Anthony et al., 2022). After identifying Symbiodinaceae cells, relative fluorescent intensity (RFI) of each cell was measured and averaged for the community. Two excitation/emission profiles were used to quantify red fluorescence. RED-B measured red emission (695/50 nm) off the blue (488 nm) excitation laser, while RED-R measured red RFI (661/15 nm) off the red (642 nm) excitation laser.
Symbiodiniaceae possess two major light harvesting complex (LHC) antennae, the peridinin-chlorophyll a protein complex (PCP) and the chlorophyll a-chlorophyll c2-peridinin protein complex (acpPC), which use peridinin and chlorophyll c2 (chl c2) as the primary light-harvesting pigments (Polívkaa et al. 2007; Hiller et al. 1993). Red light emission caused by excitation with a blue laser (RED-B) represents peridinin in the PCP or acpPC LHC antenna complex (Jiang et al. 2012; Bujak et al. 2009). Red light emission caused by excitation with the red laser (RED-R) represents the other antennae components, as a hybrid emission/excitation signature of chlorophyll a and chlorophyll c2 (chl a & chl c2) (Niedzwiedzki et al. 2014; Yacobi 2012; Zapata et al. 2001). The combination of both red fluorescent (RED-B and RED-R) signatures are referred to as ‘LHC pigments’ herein.
Green fluorescence (525/30 nm) was measured off the blue (488 nm) excitation laser (GRN-B) and was identified as a hybrid signature of beta-carotene (Lee et al., 2019; Alwis et al., 2015; Lee et al., 2012; Kleinegris et al., 2010), xanthophylls (diadinoxanthin and diatoxanthin) (Kagatani et al. 2022; Frank et al. 1996), and flavin-based fluorescent proteins (FbFPs) (Mukherjee et al. 2013; Koziol et al. 2007; Fujita et al. 2005). Given the role of beta-carotene as an efficient antioxidant (Burton, 1990), the xanthophyll cycle’s concurrence with antioxidant production (Smerilli et al. 2016), and FbFPs association with antioxidant accumulation (Deng et al. 2013; Taheri and Tarighi 2010; Sandoval et al. 2008), we refer to the green fluorescent signature (GRN-B) as ‘antioxidant pigments’.
Statistical Analysis
Symbiodiniaceae community structure was compared across sites (North, Northwest, West, South, East) and seasons (May, August). Modelled after Eckert et al. (2020) using SymPortal relative ITS2 sequence abundances and type profiles that were normalized and visualized. Multivariate homogeneity of dispersion (PERMDISP), pairwise permutation tests, and a permutational multivariate analysis of variance (PERMANOVA) were conducted on normalized ITS2 type profiles using Vegan v2.5-7 (Oksanen et al. 2020) and pairwise Adonis v0.4 (Martinez Arbizu 2017) packages. PERMDISP used Bray-Curtis similarity and permutation tests used Bray-Curtis dissimilarity. Permutation tests were run with 9999 replicates.
Physiological data violated the assumption of parametric tests of a normal data distribution, as determined by Shapiro-Wilk tests. Therefore, to evaluate factorial contributions to Symbiodiniaceae physiology, repeated measures, multivariate analyses of variance and univariate analyses of variance (RM-MANOVA & RM-ANOVA) were performed to evaluate the response of cell density, chlorophylls, peridinin, and antioxidants using the MANOVA.RM package v0.5.3 (Friedrich et al. 2022). This test neither assumes multivariate normality nor covariance matrix specificity, making it robust to repeated measure designs with factorial nesting (Friedrich et al. 2019). A repeated measures MANOVA was performed on all factorial levels [Time*Site*Plot] with a multivariate combination of cell density, peridinin, chlorophylls, and antioxidant pigments. Univariate RM-ANOVAs were performed to evaluate factorial contributions for each physiological characteristic. Main and interaction effects were resampled with 1000 non-parametric bootstrap replicates and corrected p-values were generated for modified ANOVA-type statistics. To manually compare distribution means of interest across time and site, non-parametric Kruskal-Wallis tests (X2) were performed using the FSA package v0.9.3 (Ogle et al. 2022). Given that chl a, chl c2, and peridinin co-occur within major LHCs, linear regressions were generated to compare trends in chlorophyll and peridinin fluorescence. Principal component analysis (PCA) was used to visualize the overall data structure and further reveal sources of variation. The reduced-dimension data structure was visualized using ggbiplot v0.55 (Vu, 2011). All statistical analyses were completed with R v4.1.2 (R Core Team 2021) in RStudio v1.3.1073 (RStudio Team 2020). Figures were generated and modified with a combination of ggplot2 v3.3.5 (Wickham 2016), ggpubr v0.4.0 (Kassambara 2020), and InkScape v1.1 (https://inkscape.org).
Results
Environment
Wave energy on Guam was highest from December to March with waves, on average, coming from the East year-round (Figure 2a), the island’s windward side. In 2021, Guam did not enter a coral bleaching warning (NOAA Coral Reef Watch 2019), and no bleaching was observed or reported. Therefore, 2021 represents a year characterized by natural seasonal fluctuations. Water temperature increased steadily from March to June, remaining stable during the following four months. Precipitation followed a similar trend (Figure 2b). May represented a seasonal transition with warming waters and decreasing wave energy; August was characterized by environmental stasis with high water temperatures and low wave energy (Figure 2b).
a) Spectral polar plots of aggregated historical wave data from Ritidian (red lines) and Ipan (blue lines) wave buoys. Monthly mean wave direction (black lines) indicated prevailing swells from the East, the windward side of Guam. Waves were higher with shorter periods from January to April, the season that sees strong westerly trade winds. (Provided by PacIOOS (www.pacioos.org)). b) Average sea surface temperature (SST) (spectral line) (NOAA Coral Reef Watch, 2019), and precipitation (grey bars) (Menne et al., 2012a; Menne et al., 2012b) for 2021 showed distinct seasonal patterns. The first set of samples was collected in the first two weeks of May (1) during the transitional warming period, while the second set of samples was collected in the first two weeks of August (2) during the hot, rainy season.
Symbiodiniaceae community
ITS2 metabarcoding yielded at least six replicates per site for each time point. However, plot replication was sometimes limited to a single sample, thus limiting interpretations to site level trends. Symbiodiniaceae communities of A. pulchra were largely dominated by Cladocopium C40 (Figure 3a). Beta-diversity dispersion was determined by site and not by time (Site: F = 11.725, p < 0.001; Time: F = 0.0141; p = 0.9056); biodiversity was stable with time (F = 0.437, R2 = 0.00489, p = 0.7612) (Figure 3a). Instead, Symbiodiniaceae communities were structured by site (F = 88.793, R2 = 88.793, p = 0.001) (Figure 3b). Pairwise permutation tests revealed North as an outlier, as it was the only site with a Durusdinium ITS2 type profile. A pairwise permutation test of the North across seasons indicated that communities were not statistically differentiated between sampling time points (F = 0.5426815, R2 = 0.4945, p = 0.4945). However, plot-specific data suggested Cladocopium/Durusdinium partitioning when comparing nearshore to farshore colonies, with Durusdinium being more common nearshore. ITS2 type profiles showed Symbiodiniaceae community overlap along Guam’s western coast, while southern and eastern IT2 type profiles were distinct.
a) ITS2 type diversity for endosymbiotic Symbiodiniaceae revealed high community similarity across sites and seasons. Cladocopium C40 dominated all sites except for North. b) ITS2 type profiles differed between sites, with low abundance Symbiodiniaceae clades causing differences among profiles. ITS2 type profiles did not shift significantly over time, but were distinct between sites.
Symbiodiniaceae abundance and physiology
Using cell density, peridinin, chlorophylls, and antioxidants as multivariate components, season (RM-MANOVA: t = 99.670, p < 0.001) and site (RM-MANOVA: t = 298.965, p = 0.001) were the strongest contributors to variation (Table 1). Cell density only varied with time (t = 291.668, p < 0.001), though this was solely driven by a cell density shift at the southern site. LHC pigments were heavily influenced by time (t = 291.668, p < 0.001) and site (t = 291.668, p < 0.001); antioxidant pigments only by site (t = 491.741, p < 0.001) (Table 1). The North site showed signs of nearshore to farshore ITS2-type partitioning. Therefore, plots were compared to evaluate if ITS2-type profiles were associated with functional differences. Plots neither showed differences in cell density (χ2 = 7.2067, df = 3, p = 0.06559) nor photophysiology (Flav: χ2 = 4.7533, df = 3, p-value = 0.1908; Chl: χ2 = 1.2067, df = 3, Car: p-value = 0.7514; χ2 = 1.7933, df = 3, p-value = 0.6164). As such, ITS2 type-profiles were not likely to be relevant contributors to physiological data variation.
Results of repeated measures (RM) MANOVAs to estimate the factorial contributions site and time (season) to Symbiodiniaceae cell densities and relative fluorescence intensities of peridinin, chlorophylls, and antioxidants. Top: RM-MANOVA for combined effects on cell density, photo-pigments, and antioxdiants. Bottom: Effect of site and time on individual parameters (cell density, peridinin, chlorophylls, antioxidants).
East coral colonies had much lower antioxidant and LHC pigments in May, but then converged with island-wide patterns by August (Figure 4b-e). Linear models of chlorophyll and peridinin autofluorescence confirmed a strong temporal, seasonal relationship (Figure 4e). One linear model best fit May corals (R2 = 0.765; p < 0.001), while the other linear model predicted August corals (R2 = 0.555; p < 0.001). An all encompassing linear model was still statistically supported (R2 = 0.457; p < 0.001) of the variation. However, both timepoint-specific models had a better fit (R2) to the data. Congruent with these linear models, pigments separated samples along the same PCA axis (PC1) that explained 55.9% of the variation in the data (Figure 4f). Separation of sites within time was apparent along PC1 (e.g. May:East, August:East, May: South, etc.) suggesting that photophysiology is influenced by both site and season. Cell density was the major factor of PC2, explaining 24.6% of the data variation (Figure 4f). However, no factorial levels (Site, Time, nor Plot) structured along this axis, suggesting that variation in cell density is a result of natural colony-level variation rather than environmental differences between sites or seasons.
Cell density (a) was stable across sites during both May (blue) and August (orange). Peridinin (b), chl a & chl c2 (c), and antioxidants (d) varied across sites and seasons. The East site displayed reciprocal autofluorescent signatures compared to the other sites. In May, East site Acropora pulchra displayed a visible green hue likely caused by green fluorescent protein (GFP) expression in the coral host, which was not not observed at other sites or at other times. e) Peridinin and chlorophyll autoflueorescence covaried due to their association as part of the peridinin-chlorophyll a protein complex (PCP) and the chlorophyll a-chlorophyll c2-peridinin protein complex (acpPC) in the ligh harvesting complex (LHC) antennae. Shifts in slope across time may suggest a shift in LHC reorganization as a mechanism of seasonal acclimation. All three linear regressions were significant (p < 0.001). f) Principal Component Analaysis (PCA) showed that all photophysiological signatures measured explained the majority of the variation in the data variation (55.9%), separating samples along the first axis (PC1). Cell density explained 24.6% of data variation along the second axis (PC2). Sites nested within time (ellipses) were separated along PC1 while PC2 did not yield clusters that could be related to site or time, suggesting that photophysiology was the main mode of seasonal acclimation and variation in cell density observed represents natural variation between coral colonies.
Discussion
We found that Acropora pulchra maintains a stable Symbiodiniaceae community across seasons from May, a transitional period with increasing water temperatures and decreasing wave energy, to August, a period of environmental stability with consistently high water temperatures and low wave action. Coral-dinoflagellate community acclimation is often associated with symbiont shuffling (changes in cell density or community composition) (Jones et al. 2008; Baker 2003), but we found limited evidence for seasonal symbiont shuffling. Instead, Symbiodiniaceae showed variable physiological profiles in May, a likely signal of environmental heterogeneity across sites during this period of seasonal change. Interestingly, physiological profiles converged across sites in August during environmental stasis with consistently elevated temperatures and reduced wave action (Figure 4). Given our observations, we presume that the the primary mode of seasonal acclimation in A. pulchra is Symbiodiniaceae photosystem regulation.
High community stability
We found that A. pulchra-associated Symbiodiniaceae communities were geographically structured and temporally stable (Figure 3). Long-term monitoring of endosymbiotic communities has described high stability and colony-level specificity (Rouze et al. 2019). However, this high fidelity in A. pulchra is surprising. In Acropora, Symbiodiniaceae are acquired from the environment (Baird et al. 2009), suggesting a more flexible and diverse assemblage compared to other coral genera (e.g. Qin et al. 2019; Rouzé et al. 2017). In contrast, coral colonies sampled by us were almost exclusively dominated by Cladocopium C40, with only one site containing Durusdinium D1 (Figure 3). C40 and D1 represent important lineages associated with reduced coral bleaching rates and increased coral survival following stress (Qin et al. 2019; Rouzé et al. 2017; Mieog et al. 2009; Jones et al. 2008). The co-occurrence of D1-dominated and C40-dominated ITS2 type profiles within the North site (a narrow, shallow reef flat) suggests functional similarity. However, high thermotolerance in Durusdinium is associated with trade-offs, including decreased carbon and nitrogen fixation (Silverstein et al. 2017; Keshavmurthy et al. 2014; Stat et al. 2008), and increased volatile organic compound production (Lawson et al. 2019).
The dominance of Cladocopium C40 and Durusdinium D1 in A. pulchra on Guam’s dynamic reef flats may be the result of selection, perhaps a result of recent environmentally driven mass coral mortality events (Raymundo et al. 2019; Raymundo et al. 2017). The original acquisition of C40 and D1 lineages could have occurred during coral larval settlement, increasing the likelihood of their survival during stress events (Suzuki et al. 2013), or the present A. pulchra community is composed of genets that successfully switched symbiont communities in response to extreme environmental stress (Buddemeier and Fautin 1993). Either way, Acropora-associated Symbiodiniaceae communities are determined by environmental filtering, and it is unlikely that symbiont shuffling is a mechanism for seasonal acclimation.
Photosystem dynamics
Peridinin RFI decreases from May to August in all colonies (Figure 4b) indicating an overall reduction of photopigments, a likely mechanism for acclimation to changing SSTs and photo-protection. Conversely, the incorporation of chlorophyll and antioxidant pigment signatures revealed divergent photophysiological profiles in May (Figure 4), suggesting variable acclimation responses to changing environmental conditions. The near-convergence of antioxidant and LHC profiles across sites in August (Figure 4) is likely a sign of acclimation after environmental stabilization during the wet season that led to uniformly calm and warm waters around Guam. Similar dynamics have been described in many plants and algae (e.g. Bag et al. 2020; Orefice et al. 2016; Kolari et al. 2014; Porcar-Castell et al. 2008; Dietzel et al. 2008; Johnsen et al. 1994), which indicates that Symbiodiniacae use a common mechanism of seasonal photosystem regulation in hospite.
In May, windward colonies (East) displayed a reciprocal physiological profile compared to other sites. Despite this, photophysiology converged on leeward colonies by August (Figure 4). Interestingly, windward, East coral colonies displayed a distinctive green hue in May, presumably caused by green fluorescent protein (GFP) which was absent in August and not observed at other sites (Figure 4). Coral-host GFP expression has been linked to stress mitigation through shading of photosymbionts (Lyndby et al. 2016) and antioxidant activity (Palmer et al. 2009). A. pulchra occurs in a limited area near the East site’s algal ridge, and is the only major population on the windward side of Guam. From December to May, the windward side experienced a high frequency of large swells (Figure 2a), which likely improved gas exchange (Finelli et al. 2006) and increased the availability of inorganic carbon, conditions that would increase the photosynthetic rates of coral endosymbionts (Dennison & Barnes 1988). However, if the coral host does not regulate its Symbiodiniaceae endosymbionts, the relationship can switch from symbiotic to parasitic (Morris et al. 2019). The visible expression of GFP coincided with low per cell antioxidant and LHC pigment RFIs (Figure 4b-d), suggesting that the coral host was regulating Symbiodiniaceae photophysiology (presumably with GFP) to avoid a breakdown of symbiosis. In late June, wave action decreased while SSTs increased (Figure 2). Following warming waters, decreased water flow, and increased rainfall, such host-mediated symbiont photosystem regulation would be unnecessary, explaining convergence of East colonies with island-wide patterns of Symbiodiniaceae photophysiology.
Conclusions
We found limited evidence of seasonal symbiont shuffling (Figure 3, 4a). Instead, photosystem plasticity facilitated the acclimation of Symbiodiniaceae communities to seasonal environmental change across multiple reef sites (Figure 4b-f). In light of the physiological dynamics observed here, we posit that symbiont shuffling is an acclimation strategy to extreme or unpredictable change (acute stress), while photophysiological acclimation of existing Symbiodiniaceae is used under milder or predictable environmental change (chronic stress) such as changing seasonal conditions. Investigating photoacclimation dynamics of Symbiodiniaceae in situ in conjunction with host regulatory processes (e.g. ROS scavenging) will be an important next step to elucidate the dynamic and integrated nature of photoprotective mechanisms of the coral host and Symbiodiniaceae community. Additionally, we provide a foundation for identifying the ecophysiological response norms of Symbiodiniaceae acclimation to a changing environment, prior to the eventual breakdown of coral-dinoflagellate symbiosis under severe stress such as conditions of coral bleaching.
Author Contributions
CJA and BB conceived the ideas; CJA, BMT, and BB designed the experiment; CJA, CL, and BB collected field samples; CJA and CL developed lab protocols and generated data; CJA, CL, and BMT analysed the data; CJA led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.
Acknowledgments
We wish to thank Dr. Cheryl Ames, Dr. Héloïse Rouzé, and Dr. Shinichiro Maruyama for providing their expertise and mentorship. We would also like to thank Marine Laboratory boat captain, Jonathan “Nanny” Perez, for assisting with field collections in Cocos Lagoon, and the Ritidian Eco Beach Resort for allowing land access to the pristine reef flats of Urunao. Wave data was provided by the Pacific Islands Ocean Observing System (PacIOOS) (www.pacioos.org), which is a part of the U.S. Integrated Ocean Observing System (IOOS), funded in part by National Oceanic and Atmospheric Administration (NOAA) Awards #NA16NOS0120024 and #NA21NOS0120091. Guam NSF EPSCoR directly supported this work through National Science Foundation award OIA-1946352.